721855 - Honour gfx.canvas.azure.prefer-skia pref on OS X

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[George Wright (:gw280) (needinfo me!)]

We should honour that pref on OS X. Currently it's hard coded to use CG.

Comment 1

[George Wright (:gw280) (needinfo me!)]

Attachment: Use pref on os x

Comment 2

[Jeff Muizelaar [:jrmuizel]]

Comment on attachment 592239 [details] [diff] [review]
Use pref on os x

Use GetBoolVarCache?

Comment 3

[Jeff Muizelaar [:jrmuizel]]

AddBoolVarCache

Comment 4

[George Wright (:gw280) (needinfo me!)]

Attachment: updated patch

Comment 5

[Jeff Muizelaar [:jrmuizel]]

Comment on attachment 592251 [details] [diff] [review]
updated patch

This would be better off in gfxPlatform and can be initialized in gfxPlatform's constructor.

Comment 6

[George Wright (:gw280) (needinfo me!)]

Attachment: refactor the wooooorld

Comment 7

[Joe Drew (not getting mail)]

Comment on attachment 592264 [details] [diff] [review]
refactor the wooooorld

Review of attachment 592264 [details] [diff] [review]:
-----------------------------------------------------------------

::: gfx/thebes/gfxD2DSurface.cpp
@@ +127,5 @@
>  const gfxIntSize gfxD2DSurface::GetSize() const
>  { 
>      return gfxIntSize(cairo_d2d_surface_get_width(mSurface),
>                        cairo_d2d_surface_get_height(mSurface));
> +}

THIS IS NOT PERMITTED IN THIS PATCH

::: gfx/thebes/gfxPlatform.h
@@ +449,5 @@
>  
>      // which scripts should be shaped with harfbuzz
>      PRInt32 mUseHarfBuzzScripts;
>  
> +    // The preferred azure backend to use

_DRAW TARGET_ backend

::: gfx/thebes/gfxWindowsPlatform.cpp
@@ +559,3 @@
>    }
> +
> +  return true;

Put the return with the aBackend assignment, and return false outside the if.

Comment 8

[George Wright (:gw280) (needinfo me!)]

Attachment: finally final patch

Pushed to try at https://tbpl.mozilla.org/?tree=Try&rev=28888d108556

Comment 9

[George Wright (:gw280) (needinfo me!)]

Looks like this can be checked in. The TBPL shows that the only test failures are either known intermittent failures, or are already/also present on other completely unrelated tests that have also been pushed to try.

Comment 10

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/68b7ab93105f

Comment 11

[Marco Bonardo [:mak]]

https://hg.mozilla.org/mozilla-central/rev/68b7ab93105f